Framework providing pythonic APIs, algorithms and utilities to be used with Modulus core to physics inform model training as well as higher level abstraction for domain experts
When using a numpy array based quantity (eg. x = quantity(np.ndarray([[0.2]]),'m') the call to ndim fails due to using dimensionality to check that the dimensions have been reduced. This is becuase the ndim method uses an in-place division to non-dimensionalize the quantity. An in-place operation doesn't change the _dimensionality parameter of the quantity object in pint.
I recommend using PlainQuantity.dimensionless or PlainQuantity.unitless to check that the object has been non-dimensionalized. This will work for numpy array based quantities.
Or the operation in NonDimensionalizer.ndim() could be changed to not be an in-place division.
import numpy as np
from modulus.sym import quantity
from modulus.sym.eq.non_dim import NonDimensionalizer, Scaler
# create a numpy array based quantity
x = quantity(np.array([[20.0]],"m")
print(x.dimensionality) # returns {'[length]':1.0}
# divide by 1m to get a dimensionless quantity using in-place operation
# replicates the function x.ndim()
x /= quantity(1,"m")
print(x.dimensionality) # still returns {'[length]':1.0}
print(x.dimensionless) # returns True
Version
1.2.0
On which installation method(s) does this occur?
Docker
Describe the issue
When using a numpy array based quantity (eg.
x = quantity(np.ndarray([[0.2]]),'m')the call tondimfails due to usingdimensionalityto check that the dimensions have been reduced. This is becuase thendimmethod uses an in-place division to non-dimensionalize the quantity. An in-place operation doesn't change the_dimensionalityparameter of the quantity object in pint.I recommend using
PlainQuantity.dimensionlessorPlainQuantity.unitlessto check that the object has been non-dimensionalized. This will work for numpy array based quantities.Or the operation in
NonDimensionalizer.ndim()could be changed to not be an in-place division.See related pint issue.(https://github.com/hgrecco/pint/issues/1938)
Minimum reproducible example
Relevant log output
No response
Environment details
No response
Other/Misc.
No response